Folks, go back 35 years, develop what we did with the tools that we had on the systems that we used and tell me you wouldn’t end up doing something similar!" Fortunately, this situation would eventually improve for Paul and his colleagues. "We did what we had to do back in the day. That sounds like a fairly awkward way to work, but Paul just sees it as a product of the circumstances at the time. Yes folks, these were the days you measured code performance in inches!" For performance testing you normally had some visual change on the screen (offset a screen horizontal scroll register, or alter a colour palette) at the start of a function, and then reset it at the end so that you could 'measure' how long it took on-screen by drawing marks with marker pens around this visible indicator, and seeing if the marks got closer together as you continued to optimise your code. "There were tricks for finding out where your code got to before things went wrong, like writing values into a memory location and seeing which one was written last before it crashed, then using a bisector search to narrow down to the offending code. "There were no debug tools as you’d know them today," Paul continues. If you’re a programmer wincing at that, it gets worse. As for documentation we just had the standard NES manual, which had a few typos in that needed to be corrected otherwise things wouldn't work, and also your handy 6502 instruction set guide for occasional reference, especially if you wanted to write self-modifying code." Every time you assembled your code – it was all done in one go, we didn't have linkers yet – the full binary would copy over the ribbon cable to the interface card and you'd run your code on the target machine. The PCs were Amstrads with 20MB (not a typo) hard disks. We had home-built interface cards that plugged into the cartridge slot on the NES, and they were attached by a ribbon cable to our PCs that ran PDS," he explains. "Originally at Rare everyone used PDS (Programmers Development System) which was a text editor and code assembler. The development environment that Paul worked with was rather basic. Develop what we did with the tools that we had on the systems that we used and tell me you wouldn't end up doing something similar.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |